SlideShare ist ein Scribd-Unternehmen logo
1 von 7
Downloaden Sie, um offline zu lesen
International Association of Scientific Innovation and Research (IASIR) 
(An Association Unifying the Sciences, Engineering, and Applied Research) 
International Journal of Emerging Technologies in Computational 
and Applied Sciences (IJETCAS) 
www.iasir.net 
IJETCAS 14-536; © 2014, IJETCAS All Rights Reserved Page 99 
ISSN (Print): 2279-0047 
ISSN (Online): 2279-0055 
Estimation of Inputs for a Desired Output of a Cooperative and Supportive Neural Network 
1P. Raja Sekhara Rao, 2K. Venkata Ratnam, 3P.Lalitha 
1Department of Mathematics, Government Polytechnic, Addanki - 523 201, Prakasam, A.P., INDIA. 
2Department of Mathematics, Birla Institute of Technology and Science-Pilani, Hyderabad campus, 
Jawahar Nagar, Hyderabad-500078, INDIA. 
3Department of Mathematics, St. Francis College for Women, Begumpet, Hyderabad, INDIA. 
__________________________________________________________________________________________ 
Abstract: In this paper a cooperative and supportive neural network involving time delays is considered. External inputs to the network are allowed to vary with respect to time. Asymptotic behavior of solutions of network system with variable inputs is studied with respect to its counterpart of constant inputs. With suitable restrictions on the inputs, it is noticed that solution of the network may be made to approach a pre-specified output. 
Keywords: Co-operative and Supportive Neural Network, Variable Inputs, Desired Output, Convergence. 
__________________________________________________________________________________________ 
I. Introduction 
This paper deals with the study of influence of time varying exogenous inputs on a cooperative and supportive neural network. A model of a cooperative and supportive network (CSNN, for short) is introduced by Sree Hari Rao and Raja Sekhara Rao [9]. It takes into account the collective capabilities of neurons involved with tasks divided and distributed to sub networks of neurons. Applications to such networks are many, for example, in industrial information management (hierarchical systems) which involve distribution and monitoring of various tasks. They are also useful in classification and clustering problems, data mining and financial engineering [6,7,8]. They are also utilized for parameter estimation of auto regressive signals and to decompose complex classification tasks into simpler subtasks and solve then. 
In a recent paper [11], the authors considered time delays in transmission of information from sub-networks to main one as well as in processing of information in sub-network itself (before transmission of information to main network). Qualitative properties of solutions of the system are studied. Sufficient conditions for global asymptotic stability of equilibrium pattern of the system are established even in the presence of time delays. In the present paper, we wish to consider the CSNN model of [11] with time delays to study the influence of time varying inputs on the system. The motivation for this study stems from the observations of [10] that the applicability a neural network may be increased by the choice of inputs and inputs play a key role in attaining desired outputs. Proper choice of could be an alternative for modifying the neural network for each application and existing neural network may be utilized for different tasks, thus. Besides this, the presence of time varying inputs make the system non autonomous and the study enriches the literature. Mathematical studies of neural networks have been concentrated on stability of equilibrium patterns. Equilibria are stationary solutions of the system and correspond to memory states of the network. Stability of an equilibrium implies a recall of memory state. Thus, such stability analysis of neural networks is confined to recall of memories only and we may not reach the desired output for which the network is intended. In the present study, we deviate from this recall of memories but look for ways of reaching a desired solution. 
An attempt is made in [9] to explain briefly the influence of variable inputs on asymptotic nature of solutions of CSNN model. The present study extends this work. We concentrate on the interplay between the inputs and outputs of the network. For this, several results are established for estimation or restriction of inputs for getting a desired or pre-specified output and understand the behavior of solutions in the presence of variable inputs. The work also extends the study of [10] carried out for BAM networks. As remarked in [10], convergence to a desired output for a given output explained here should not be confused with convergence of output function of the network. Results are available in literature which consider time varying inputs in various directions [1- 3,5,12] but our emphasis here is on utilization of these inputs to make solutions of system approach an a priori value of output. We reiterate that this is not yet another usual study on qualitative behavior of solutions of the system under the influence of variable inputs. 
The paper is organized as follows. In Section 2, the model under consideration is explained. Asymptotic behavior of solutions and their relation with the solutions of corresponding system with constant inputs are discussed. Section 3 deals with the input-output trade-off. Estimates on inputs are provided for approaching a desired, preset output for the network. A discussion follows in Section 4.
P. Raja Sekhara Rao et al., International Journal of Emerging Technologies in Computational and Applied Sciences, 9(1), June-August, 
2014,pp. 99-105 
IJETCAS 14-536; © 2014, IJETCAS All Rights Reserved Page 100 
II. The Model and Asymptotic Behavior 
The following model is considered in [11], 
, 
In (2.1), , i=1,2,...,n denote a typical neuron in neural field X and denote a subgroup of neurons in another neuronal field Y and are attached to . may be considered to form the main group of neurons which are required to perform the task. constitute a subgroup of neurons attached to each to 
which assigns some of its task. support, coordinate and cooperate with in completing the task. and are positive constants known as passive decay rates of neurons and respectively. and are the synaptic connection weights and all these are assumed to be real or complex constants. denotes the rate of distribution of information between and . The weight connections connect the i th neuron in one neuronal field to the j th neuron in another neuronal field. The functions and are the neuronal output response functions and are more commonly known as the signal functions. The parameter signifies the time delay in transmission of information from sub network neuron to main network neuron . The delay in second equation represents the processing delays in the subsystems. and are exogenous inputs which are assumed to be constants in [11]. For more details of the terms and design of the CSNN, readers are referred to [9]. 
Introducing the time variables and (t), , in place of constant inputs and into the system (2.1), we get 
The following initial functions are assumed for the system (2.2). 
for , (2.3) 
where are continuous, bounded functions on and We assume that the response functions and satisfy conditions 
(2.4) 
(2.5) 
(2.6) 
where , , and are positive constants. 
Under the conditions (2.4)-(2.6) on the response functions and with and bounded, continuous functions on ) it is not difficult to see that the system (2.2) possesses unique solutions that are continuable in their maximal intervals of existence ([11]). 
Since (2.2) is non-autonomous it may not possess equilibrium patterns(constant solutions). A solution of (2.1) or (2.2) is denoted by where , throughout. 
Therefore, we study the asymptotic behavior of its solutions. We recall from [10] that two solutions and of the system (2.2) are asymptotically near if 
. 
In the following, we present results on asymptotic nearness of solutions of (2.2). 
Our first result is 
Theorem 2.1: For any pair of solutions and of (2.2), we have 
provided holds, where 
(2.7)
P. Raja Sekhara Rao et al., International Journal of Emerging Technologies in Computational and Applied Sciences, 9(1), June-August, 
2014,pp. 99-105 
IJETCAS 14-536; © 2014, IJETCAS All Rights Reserved Page 101 
Proof: Consider the functional, 
Along the solutions of (2.2), the upper right derivative of V is given by 
using (2.4)-(2.6). 
. 
Integrating both sides with respect to t, 
Thus, V (t) is bounded on and for . But and are also bounded on . Hence, it follows that their derivatives are also bounded on
P. Raja Sekhara Rao et al., International Journal of Emerging Technologies in Computational and Applied Sciences, 9(1), June-August, 
2014,pp. 99-105 
IJETCAS 14-536; © 2014, IJETCAS All Rights Reserved Page 102 
Therefore, and , are uniformly continuous on . Thus, we may conclude that and (e.g., [4]). This concludes the proof. 
The following result provides conditions under which all solutions of (2.2) are asymptotic to the solutions of (2.1), which shows that for a proper choice of input functions the stability of system (2.1) is not altered by the presence of time dependent inputs. 
Theorem 2.2: Assume that the parametric conditions (2.7) hold. Further, let inputs satisfy where . Then for any solutions (x, y) of (2.2) and 
Proof: To establish this, we employ the same functional as in Theorem 2.1, that is, 
Proceeding as in Theorem 2.1 we get after a simplification and rearrangement 
+ 
Rest of the argument is same as that of Theorems 2.1, and hence, omitted. Thus, the conclusion follows. 
We now recall from [11] that the system (2.1) has a unique equilibrium pattern any set of input vectors provided the parameters satisfy, 
Then we have, 
Corollary 2.3: Assume that all the hypotheses of Theorem 2.2 are satisfied. Further if (2.1) possesses equilibrium pattern then all solutions (x, y) of (2.2) approach 
Proof: The result obviously follows form the observation that the equilibrium pattern is also a solution of (2.1) and the choice in Theorem 2.2. 
The following example illustrates the above results. 
Example 2.4: Consider the following system having two neurons in X each supported by two neurons in Y involving time delays as given by 
. 
Choose and . Then . 
Clearly conditions for both the existence of unique equilibrium and its stability are satisfied for any pair of constant inputs .
P. Raja Sekhara Rao et al., International Journal of Emerging Technologies in Computational and Applied Sciences, 9(1), June-August, 
2014,pp. 99-105 
IJETCAS 14-536; © 2014, IJETCAS All Rights Reserved Page 103 
Now choose 
It is easy to see that since the condition holds, and we have 
(i). Conditions of Theorem 2.1 are satisfied and all solutions of the system are asymptotic to each other. 
(ii). Conditions of Corollary 2.3 are satisfied and all solutions of the system approach the equilibrium pattern of corresponding system with constant inputs. 
III. Estimations on Inputs for a Pre-specified Output 
In this section, we try to estimate our inputs, depending on the output given, that help the solutions to approach the given output. For an easy understanding of the concept, we avoid the complicated notation. We, therefore, rearrange our system (2.2) suitably. We use the following notation 
For , (2.2) may be represented as 
(3.1) 
in which 
We assume that is the desired output of the network. Further that both and are fixed with respect to t and are arbitrarily chosen. We now arrange (3.1) as 
(3.2) 
Conditions on response functions (2.2) to (2.6) may be modified as 
and for some || 
denoting appropriate norm. We have 
Theorem 3.1. Assume that the parameters of the system and the response functions satisfy the condition 
For an arbitrarily chosen output , the solutions of system (3.1) converge to provided the inputs satisfy either of the conditions 
or 
Proof . Then the upper right derivative of V along the solutions of (3.1), using (3.2), we have 
– – – 
This gives rise to 
Rest of the argument is similar to that of Theorem 2.1 and invoking the condition (i) on inputs. 
Again, it is easy to see from the last inequality above that 
for large t using conditions (ii) on and . Hence, in either of the cases, follows. 
The proof is complete. 
The following example illustrates the effectiveness of this result.
P. Raja Sekhara Rao et al., International Journal of Emerging Technologies in Computational and Applied Sciences, 9(1), June-August, 
2014,pp. 99-105 
IJETCAS 14-536; © 2014, IJETCAS All Rights Reserved Page 104 
Example 3.2: Consider 
Now choose and 
Clearly, for given we have all the conditions of Theorem 3.1 are satisfied, and hence, for sufficiently large t. 
We shall now consider the time delay system corresponding to (2.2), 
(3.3) 
As has been done earlier for a given , we write (3.3) as 
(3.4) 
We have 
Theorem 3.3: Assume that the parameters of the system and the response functions satisfy the condition 
For an arbitrarily chosen output , solutions of system (3.3) converge to provided the inputs satisfy the condition 
Proof: Employing the functional 
using (3.4) and proceeding as in Theorem 2.1 and Theorem 3.1, the conclusion follows. 
Since the conditions on parameters and input functions in Theorem 3.1 and 3.3 are the same, it may imply that delays have no effect on convergence here 
Example 3.4: Consider 
Now choose and 
Clearly, for given we have all the conditions of Theorem 3.1 are satisfied, and hence, for sufficiently large t. 
Remark 3.5: Now consider the system 
(3.5) 
are constant inputs, given . It is easy to observe that is an equilibrium pattern of (3.5). Then from Corollary 2.4, we have, solutions of (3.3) approach whenever the variable inputs of (3.3) are well near those of (3.5) as specified in Corollary 2.4. 
Thus, by varying the external inputs of the system in the parameter space defined by as specified by Theorems 3.1 and 3.3, the solutions of the network approaches pre-specified output 
IV. Discussion 
In this article, we have extended the concept of approaching a desired output of a given network by suitable selection of inputs based on the given output for a cooperative and supportive neural network that was studied earlier for a BAM network ([10]). With the help of suitable Lyapunov functionals, results are established for asymptotic nearness and boundedness of solutions of the system also. It is noticed that inputs define a new space of equilibria for the network while they run through a space defined by output parameters. This way memory states of brain that are usually ignored by constant inputs may be recalled by varying the inputs to brain appropriately. Since the input-output relation is not direct but includes system parameters and functional responses, dynamics of entire system are involved in this process. It is hoped that this concept helps in utilizing the same network for different applications without altering its architecture. This shows how designed structures may be made emergent structures which are adaptive and flexible. Since the results hold good for all time delays (delay independent criteria) the results are applicable to delay-free case as well, i.e., models of [9].
P. Raja Sekhara Rao et al., International Journal of Emerging Technologies in Computational and Applied Sciences, 9(1), June-August, 
2014,pp. 99-105 
IJETCAS 14-536; © 2014, IJETCAS All Rights Reserved Page 105 
V. References 
[1] H. Bereketoglu and I. Gyori, Global asymptotic stability in a nonautonomous Lotka-Volterra type system with infinite delay, Journal of Mathematical Analysis and Applications, 210(1997), 279-291. 
[2] Q.X. Dong, K. Matsui and X.K. Huang, Existence and stability of periodic solutions for Hopfield neural network equations with periodic input, Nonlinear Analysis, 49(2002), 471-479. 
[3] M. Forti, P. Nistri and D. Papini, Global exponential stability and global convergence in finite time of delayed neural networks with infinite gain, IEEE TNN 16(6): 1449-1463, 2005. 
[4] K. Gopalsamy and Xue-Zhong He, Delay-independent stability in bidirectional associative memory networks, IEEE TNN, 5(1994), 998-1002. 
[5] S. Hu and D. Liu, On global output convergence of a class of recurrent neural networks with time varying inputs, 18(2005), 171- 178. 
[6] B. Kosko, “Neural Networks and Fuzzy Systems - A Dynamical Systems Approach to Machine Intelligence", Prentice-Hall of India, New Delhi, 1994. 
[7] F.-L. Luo, R. Unbehauen, Applied Neural Networks for Signal Processing, Cambridge Univ. Press, Cambridge, UK, 1997. 
[8] B.B. Misra and S. Dehuri, Functional link artificial neural network for classification task in data mining, J. Computer Science, 3(12), 2007, 948-955. 
[9] V. Sree Hari Rao and P. Raja Sekhara Rao, Cooperative and Supportive Neural Networks, Physics Letters A 371 (2007) 101–110. 
[10] V. Sree Hari Rao and P. Raja Sekhara Rao, Time Varying Stimulations to Simple Neural Networks and Convergence to Desired 
Outputs, Communicated. 
[11] P. Raja Sekhara Rao, K.Venkata Ratnam and P. Lalitha, Delay Independent Stability of Co-operative and Supportive Neural Networks, Communicated. 
[12] Zhang Yi, J.C. Lv and L. Zhang, Output convergence analysis for a class of delayed recurrent neural networks with time varying inputs, IEEE TSMC, 36(1), 87-95, 2006.

Weitere ähnliche Inhalte

Was ist angesagt?

Hybrid PSO-SA algorithm for training a Neural Network for Classification
Hybrid PSO-SA algorithm for training a Neural Network for ClassificationHybrid PSO-SA algorithm for training a Neural Network for Classification
Hybrid PSO-SA algorithm for training a Neural Network for Classification
IJCSEA Journal
 
Scalable Constrained Spectral Clustering
Scalable Constrained Spectral ClusteringScalable Constrained Spectral Clustering
Scalable Constrained Spectral Clustering
1crore projects
 
Model reduction-of-linear-systems-by conventional-and-evolutionary-techniques
Model reduction-of-linear-systems-by conventional-and-evolutionary-techniquesModel reduction-of-linear-systems-by conventional-and-evolutionary-techniques
Model reduction-of-linear-systems-by conventional-and-evolutionary-techniques
Cemal Ardil
 
Abstract takahashi
Abstract takahashiAbstract takahashi
Abstract takahashi
harmonylab
 

Was ist angesagt? (20)

New Design Architecture of Chaotic Secure Communication System Combined with ...
New Design Architecture of Chaotic Secure Communication System Combined with ...New Design Architecture of Chaotic Secure Communication System Combined with ...
New Design Architecture of Chaotic Secure Communication System Combined with ...
 
International Refereed Journal of Engineering and Science (IRJES)
International Refereed Journal of Engineering and Science (IRJES)International Refereed Journal of Engineering and Science (IRJES)
International Refereed Journal of Engineering and Science (IRJES)
 
Hybrid PSO-SA algorithm for training a Neural Network for Classification
Hybrid PSO-SA algorithm for training a Neural Network for ClassificationHybrid PSO-SA algorithm for training a Neural Network for Classification
Hybrid PSO-SA algorithm for training a Neural Network for Classification
 
Survey on classification algorithms for data mining (comparison and evaluation)
Survey on classification algorithms for data mining (comparison and evaluation)Survey on classification algorithms for data mining (comparison and evaluation)
Survey on classification algorithms for data mining (comparison and evaluation)
 
Probabilistic Methods Of Signal And System Analysis, 3rd Edition
Probabilistic Methods Of Signal And System Analysis, 3rd EditionProbabilistic Methods Of Signal And System Analysis, 3rd Edition
Probabilistic Methods Of Signal And System Analysis, 3rd Edition
 
IRJET- Semantics based Document Clustering
IRJET- Semantics based Document ClusteringIRJET- Semantics based Document Clustering
IRJET- Semantics based Document Clustering
 
Extended pso algorithm for improvement problems k means clustering algorithm
Extended pso algorithm for improvement problems k means clustering algorithmExtended pso algorithm for improvement problems k means clustering algorithm
Extended pso algorithm for improvement problems k means clustering algorithm
 
Different Similarity Measures for Text Classification Using Knn
Different Similarity Measures for Text Classification Using KnnDifferent Similarity Measures for Text Classification Using Knn
Different Similarity Measures for Text Classification Using Knn
 
Scalable Constrained Spectral Clustering
Scalable Constrained Spectral ClusteringScalable Constrained Spectral Clustering
Scalable Constrained Spectral Clustering
 
Cb pattern trees identifying
Cb pattern trees  identifyingCb pattern trees  identifying
Cb pattern trees identifying
 
Model reduction-of-linear-systems-by conventional-and-evolutionary-techniques
Model reduction-of-linear-systems-by conventional-and-evolutionary-techniquesModel reduction-of-linear-systems-by conventional-and-evolutionary-techniques
Model reduction-of-linear-systems-by conventional-and-evolutionary-techniques
 
Az36311316
Az36311316Az36311316
Az36311316
 
Designing of an efficient algorithm for identifying Abbreviation definitions ...
Designing of an efficient algorithm for identifying Abbreviation definitions ...Designing of an efficient algorithm for identifying Abbreviation definitions ...
Designing of an efficient algorithm for identifying Abbreviation definitions ...
 
Bragged Regression Tree Algorithm for Dynamic Distribution and Scheduling of ...
Bragged Regression Tree Algorithm for Dynamic Distribution and Scheduling of ...Bragged Regression Tree Algorithm for Dynamic Distribution and Scheduling of ...
Bragged Regression Tree Algorithm for Dynamic Distribution and Scheduling of ...
 
40220140501006
4022014050100640220140501006
40220140501006
 
Automated Essay Grading using Features Selection
Automated Essay Grading using Features SelectionAutomated Essay Grading using Features Selection
Automated Essay Grading using Features Selection
 
Classification of text data using feature clustering algorithm
Classification of text data using feature clustering algorithmClassification of text data using feature clustering algorithm
Classification of text data using feature clustering algorithm
 
A Comparative Analysis of Feature Selection Methods for Clustering DNA Sequences
A Comparative Analysis of Feature Selection Methods for Clustering DNA SequencesA Comparative Analysis of Feature Selection Methods for Clustering DNA Sequences
A Comparative Analysis of Feature Selection Methods for Clustering DNA Sequences
 
Abstract takahashi
Abstract takahashiAbstract takahashi
Abstract takahashi
 
Ip3514921495
Ip3514921495Ip3514921495
Ip3514921495
 

Andere mochten auch (18)

Ijebea14 215
Ijebea14 215Ijebea14 215
Ijebea14 215
 
Ijetcas14 392
Ijetcas14 392Ijetcas14 392
Ijetcas14 392
 
Ijetcas14 443
Ijetcas14 443Ijetcas14 443
Ijetcas14 443
 
Ijetcas14 474
Ijetcas14 474Ijetcas14 474
Ijetcas14 474
 
Ijetcas14 357
Ijetcas14 357Ijetcas14 357
Ijetcas14 357
 
Aijrfans14 269
Aijrfans14 269Aijrfans14 269
Aijrfans14 269
 
Aijrfans14 243
Aijrfans14 243Aijrfans14 243
Aijrfans14 243
 
Ijetcas14 479
Ijetcas14 479Ijetcas14 479
Ijetcas14 479
 
Ijetcas14 460
Ijetcas14 460Ijetcas14 460
Ijetcas14 460
 
Ijetcas14 435
Ijetcas14 435Ijetcas14 435
Ijetcas14 435
 
Ijetcas14 562
Ijetcas14 562Ijetcas14 562
Ijetcas14 562
 
Ijetcas14 371
Ijetcas14 371Ijetcas14 371
Ijetcas14 371
 
Ijetcas14 312
Ijetcas14 312Ijetcas14 312
Ijetcas14 312
 
Ijetcas14 315
Ijetcas14 315Ijetcas14 315
Ijetcas14 315
 
Sulfide Minerals
Sulfide MineralsSulfide Minerals
Sulfide Minerals
 
Aijrfans14 227
Aijrfans14 227Aijrfans14 227
Aijrfans14 227
 
Ijetcas14 639
Ijetcas14 639Ijetcas14 639
Ijetcas14 639
 
Ijetcas14 648
Ijetcas14 648Ijetcas14 648
Ijetcas14 648
 

Ähnlich wie Ijetcas14 536

AN IMPROVED METHOD FOR IDENTIFYING WELL-TEST INTERPRETATION MODEL BASED ON AG...
AN IMPROVED METHOD FOR IDENTIFYING WELL-TEST INTERPRETATION MODEL BASED ON AG...AN IMPROVED METHOD FOR IDENTIFYING WELL-TEST INTERPRETATION MODEL BASED ON AG...
AN IMPROVED METHOD FOR IDENTIFYING WELL-TEST INTERPRETATION MODEL BASED ON AG...
IAEME Publication
 
Urban strategies to promote resilient cities The case of enhancing Historic C...
Urban strategies to promote resilient cities The case of enhancing Historic C...Urban strategies to promote resilient cities The case of enhancing Historic C...
Urban strategies to promote resilient cities The case of enhancing Historic C...
inventionjournals
 
Analysis of intelligent system design by neuro adaptive control no restriction
Analysis of intelligent system design by neuro adaptive control no restrictionAnalysis of intelligent system design by neuro adaptive control no restriction
Analysis of intelligent system design by neuro adaptive control no restriction
iaemedu
 
Analysis of intelligent system design by neuro adaptive control
Analysis of intelligent system design by neuro adaptive controlAnalysis of intelligent system design by neuro adaptive control
Analysis of intelligent system design by neuro adaptive control
iaemedu
 

Ähnlich wie Ijetcas14 536 (20)

NEURAL NETWORK FOR THE RELIABILITY ANALYSIS OF A SERIES - PARALLEL SYSTEM SUB...
NEURAL NETWORK FOR THE RELIABILITY ANALYSIS OF A SERIES - PARALLEL SYSTEM SUB...NEURAL NETWORK FOR THE RELIABILITY ANALYSIS OF A SERIES - PARALLEL SYSTEM SUB...
NEURAL NETWORK FOR THE RELIABILITY ANALYSIS OF A SERIES - PARALLEL SYSTEM SUB...
 
A Hybrid Deep Neural Network Model For Time Series Forecasting
A Hybrid Deep Neural Network Model For Time Series ForecastingA Hybrid Deep Neural Network Model For Time Series Forecasting
A Hybrid Deep Neural Network Model For Time Series Forecasting
 
Survey on Artificial Neural Network Learning Technique Algorithms
Survey on Artificial Neural Network Learning Technique AlgorithmsSurvey on Artificial Neural Network Learning Technique Algorithms
Survey on Artificial Neural Network Learning Technique Algorithms
 
EXPERIMENTS ON DIFFERENT RECURRENT NEURAL NETWORKS FOR ENGLISH-HINDI MACHINE ...
EXPERIMENTS ON DIFFERENT RECURRENT NEURAL NETWORKS FOR ENGLISH-HINDI MACHINE ...EXPERIMENTS ON DIFFERENT RECURRENT NEURAL NETWORKS FOR ENGLISH-HINDI MACHINE ...
EXPERIMENTS ON DIFFERENT RECURRENT NEURAL NETWORKS FOR ENGLISH-HINDI MACHINE ...
 
AN IMPROVED METHOD FOR IDENTIFYING WELL-TEST INTERPRETATION MODEL BASED ON AG...
AN IMPROVED METHOD FOR IDENTIFYING WELL-TEST INTERPRETATION MODEL BASED ON AG...AN IMPROVED METHOD FOR IDENTIFYING WELL-TEST INTERPRETATION MODEL BASED ON AG...
AN IMPROVED METHOD FOR IDENTIFYING WELL-TEST INTERPRETATION MODEL BASED ON AG...
 
Urban strategies to promote resilient cities The case of enhancing Historic C...
Urban strategies to promote resilient cities The case of enhancing Historic C...Urban strategies to promote resilient cities The case of enhancing Historic C...
Urban strategies to promote resilient cities The case of enhancing Historic C...
 
Power system transient stability margin estimation using artificial neural ne...
Power system transient stability margin estimation using artificial neural ne...Power system transient stability margin estimation using artificial neural ne...
Power system transient stability margin estimation using artificial neural ne...
 
Time Series Forecasting Using Novel Feature Extraction Algorithm and Multilay...
Time Series Forecasting Using Novel Feature Extraction Algorithm and Multilay...Time Series Forecasting Using Novel Feature Extraction Algorithm and Multilay...
Time Series Forecasting Using Novel Feature Extraction Algorithm and Multilay...
 
A Learning Linguistic Teaching Control for a Multi-Area Electric Power System
A Learning Linguistic Teaching Control for a Multi-Area Electric Power SystemA Learning Linguistic Teaching Control for a Multi-Area Electric Power System
A Learning Linguistic Teaching Control for a Multi-Area Electric Power System
 
A Novel Neuroglial Architecture for Modelling Singular Perturbation System
A Novel Neuroglial Architecture for Modelling Singular Perturbation System  A Novel Neuroglial Architecture for Modelling Singular Perturbation System
A Novel Neuroglial Architecture for Modelling Singular Perturbation System
 
28 15017 estimation of turbidity in water(edit)
28 15017 estimation of turbidity in water(edit)28 15017 estimation of turbidity in water(edit)
28 15017 estimation of turbidity in water(edit)
 
Multiprocessor scheduling of dependent tasks to minimize makespan and reliabi...
Multiprocessor scheduling of dependent tasks to minimize makespan and reliabi...Multiprocessor scheduling of dependent tasks to minimize makespan and reliabi...
Multiprocessor scheduling of dependent tasks to minimize makespan and reliabi...
 
Analysis of intelligent system design by neuro adaptive control no restriction
Analysis of intelligent system design by neuro adaptive control no restrictionAnalysis of intelligent system design by neuro adaptive control no restriction
Analysis of intelligent system design by neuro adaptive control no restriction
 
Analysis of intelligent system design by neuro adaptive control
Analysis of intelligent system design by neuro adaptive controlAnalysis of intelligent system design by neuro adaptive control
Analysis of intelligent system design by neuro adaptive control
 
Kavitha soft computing
Kavitha soft computingKavitha soft computing
Kavitha soft computing
 
Optimal neural network models for wind speed prediction
Optimal neural network models for wind speed predictionOptimal neural network models for wind speed prediction
Optimal neural network models for wind speed prediction
 
Optimal neural network models for wind speed prediction
Optimal neural network models for wind speed predictionOptimal neural network models for wind speed prediction
Optimal neural network models for wind speed prediction
 
Optimal neural network models for wind speed prediction
Optimal neural network models for wind speed predictionOptimal neural network models for wind speed prediction
Optimal neural network models for wind speed prediction
 
THE ACTIVE CONTROLLER DESIGN FOR ACHIEVING GENERALIZED PROJECTIVE SYNCHRONIZA...
THE ACTIVE CONTROLLER DESIGN FOR ACHIEVING GENERALIZED PROJECTIVE SYNCHRONIZA...THE ACTIVE CONTROLLER DESIGN FOR ACHIEVING GENERALIZED PROJECTIVE SYNCHRONIZA...
THE ACTIVE CONTROLLER DESIGN FOR ACHIEVING GENERALIZED PROJECTIVE SYNCHRONIZA...
 
FUNCTION PROJECTIVE SYNCHRONIZATION OF NEW CHAOTIC REVERSAL SYSTEMS
FUNCTION PROJECTIVE SYNCHRONIZATION OF NEW CHAOTIC REVERSAL SYSTEMSFUNCTION PROJECTIVE SYNCHRONIZATION OF NEW CHAOTIC REVERSAL SYSTEMS
FUNCTION PROJECTIVE SYNCHRONIZATION OF NEW CHAOTIC REVERSAL SYSTEMS
 

Mehr von Iasir Journals

Mehr von Iasir Journals (20)

ijetcas14 650
ijetcas14 650ijetcas14 650
ijetcas14 650
 
Ijetcas14 647
Ijetcas14 647Ijetcas14 647
Ijetcas14 647
 
Ijetcas14 643
Ijetcas14 643Ijetcas14 643
Ijetcas14 643
 
Ijetcas14 641
Ijetcas14 641Ijetcas14 641
Ijetcas14 641
 
Ijetcas14 632
Ijetcas14 632Ijetcas14 632
Ijetcas14 632
 
Ijetcas14 624
Ijetcas14 624Ijetcas14 624
Ijetcas14 624
 
Ijetcas14 619
Ijetcas14 619Ijetcas14 619
Ijetcas14 619
 
Ijetcas14 615
Ijetcas14 615Ijetcas14 615
Ijetcas14 615
 
Ijetcas14 608
Ijetcas14 608Ijetcas14 608
Ijetcas14 608
 
Ijetcas14 605
Ijetcas14 605Ijetcas14 605
Ijetcas14 605
 
Ijetcas14 604
Ijetcas14 604Ijetcas14 604
Ijetcas14 604
 
Ijetcas14 598
Ijetcas14 598Ijetcas14 598
Ijetcas14 598
 
Ijetcas14 594
Ijetcas14 594Ijetcas14 594
Ijetcas14 594
 
Ijetcas14 593
Ijetcas14 593Ijetcas14 593
Ijetcas14 593
 
Ijetcas14 591
Ijetcas14 591Ijetcas14 591
Ijetcas14 591
 
Ijetcas14 589
Ijetcas14 589Ijetcas14 589
Ijetcas14 589
 
Ijetcas14 585
Ijetcas14 585Ijetcas14 585
Ijetcas14 585
 
Ijetcas14 584
Ijetcas14 584Ijetcas14 584
Ijetcas14 584
 
Ijetcas14 583
Ijetcas14 583Ijetcas14 583
Ijetcas14 583
 
Ijetcas14 580
Ijetcas14 580Ijetcas14 580
Ijetcas14 580
 

Kürzlich hochgeladen

Integrated Test Rig For HTFE-25 - Neometrix
Integrated Test Rig For HTFE-25 - NeometrixIntegrated Test Rig For HTFE-25 - Neometrix
Integrated Test Rig For HTFE-25 - Neometrix
Neometrix_Engineering_Pvt_Ltd
 
Cara Menggugurkan Sperma Yang Masuk Rahim Biyar Tidak Hamil
Cara Menggugurkan Sperma Yang Masuk Rahim Biyar Tidak HamilCara Menggugurkan Sperma Yang Masuk Rahim Biyar Tidak Hamil
Cara Menggugurkan Sperma Yang Masuk Rahim Biyar Tidak Hamil
Cara Menggugurkan Kandungan 087776558899
 
Hospital management system project report.pdf
Hospital management system project report.pdfHospital management system project report.pdf
Hospital management system project report.pdf
Kamal Acharya
 
Kuwait City MTP kit ((+919101817206)) Buy Abortion Pills Kuwait
Kuwait City MTP kit ((+919101817206)) Buy Abortion Pills KuwaitKuwait City MTP kit ((+919101817206)) Buy Abortion Pills Kuwait
Kuwait City MTP kit ((+919101817206)) Buy Abortion Pills Kuwait
jaanualu31
 
Standard vs Custom Battery Packs - Decoding the Power Play
Standard vs Custom Battery Packs - Decoding the Power PlayStandard vs Custom Battery Packs - Decoding the Power Play
Standard vs Custom Battery Packs - Decoding the Power Play
Epec Engineered Technologies
 
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
ssuser89054b
 

Kürzlich hochgeladen (20)

Integrated Test Rig For HTFE-25 - Neometrix
Integrated Test Rig For HTFE-25 - NeometrixIntegrated Test Rig For HTFE-25 - Neometrix
Integrated Test Rig For HTFE-25 - Neometrix
 
Unit 4_Part 1 CSE2001 Exception Handling and Function Template and Class Temp...
Unit 4_Part 1 CSE2001 Exception Handling and Function Template and Class Temp...Unit 4_Part 1 CSE2001 Exception Handling and Function Template and Class Temp...
Unit 4_Part 1 CSE2001 Exception Handling and Function Template and Class Temp...
 
Cara Menggugurkan Sperma Yang Masuk Rahim Biyar Tidak Hamil
Cara Menggugurkan Sperma Yang Masuk Rahim Biyar Tidak HamilCara Menggugurkan Sperma Yang Masuk Rahim Biyar Tidak Hamil
Cara Menggugurkan Sperma Yang Masuk Rahim Biyar Tidak Hamil
 
Block diagram reduction techniques in control systems.ppt
Block diagram reduction techniques in control systems.pptBlock diagram reduction techniques in control systems.ppt
Block diagram reduction techniques in control systems.ppt
 
Employee leave management system project.
Employee leave management system project.Employee leave management system project.
Employee leave management system project.
 
Generative AI or GenAI technology based PPT
Generative AI or GenAI technology based PPTGenerative AI or GenAI technology based PPT
Generative AI or GenAI technology based PPT
 
Hospital management system project report.pdf
Hospital management system project report.pdfHospital management system project report.pdf
Hospital management system project report.pdf
 
Thermal Engineering Unit - I & II . ppt
Thermal Engineering  Unit - I & II . pptThermal Engineering  Unit - I & II . ppt
Thermal Engineering Unit - I & II . ppt
 
Thermal Engineering -unit - III & IV.ppt
Thermal Engineering -unit - III & IV.pptThermal Engineering -unit - III & IV.ppt
Thermal Engineering -unit - III & IV.ppt
 
A Study of Urban Area Plan for Pabna Municipality
A Study of Urban Area Plan for Pabna MunicipalityA Study of Urban Area Plan for Pabna Municipality
A Study of Urban Area Plan for Pabna Municipality
 
HAND TOOLS USED AT ELECTRONICS WORK PRESENTED BY KOUSTAV SARKAR
HAND TOOLS USED AT ELECTRONICS WORK PRESENTED BY KOUSTAV SARKARHAND TOOLS USED AT ELECTRONICS WORK PRESENTED BY KOUSTAV SARKAR
HAND TOOLS USED AT ELECTRONICS WORK PRESENTED BY KOUSTAV SARKAR
 
Kuwait City MTP kit ((+919101817206)) Buy Abortion Pills Kuwait
Kuwait City MTP kit ((+919101817206)) Buy Abortion Pills KuwaitKuwait City MTP kit ((+919101817206)) Buy Abortion Pills Kuwait
Kuwait City MTP kit ((+919101817206)) Buy Abortion Pills Kuwait
 
COST-EFFETIVE and Energy Efficient BUILDINGS ptx
COST-EFFETIVE  and Energy Efficient BUILDINGS ptxCOST-EFFETIVE  and Energy Efficient BUILDINGS ptx
COST-EFFETIVE and Energy Efficient BUILDINGS ptx
 
data_management_and _data_science_cheat_sheet.pdf
data_management_and _data_science_cheat_sheet.pdfdata_management_and _data_science_cheat_sheet.pdf
data_management_and _data_science_cheat_sheet.pdf
 
S1S2 B.Arch MGU - HOA1&2 Module 3 -Temple Architecture of Kerala.pptx
S1S2 B.Arch MGU - HOA1&2 Module 3 -Temple Architecture of Kerala.pptxS1S2 B.Arch MGU - HOA1&2 Module 3 -Temple Architecture of Kerala.pptx
S1S2 B.Arch MGU - HOA1&2 Module 3 -Temple Architecture of Kerala.pptx
 
Standard vs Custom Battery Packs - Decoding the Power Play
Standard vs Custom Battery Packs - Decoding the Power PlayStandard vs Custom Battery Packs - Decoding the Power Play
Standard vs Custom Battery Packs - Decoding the Power Play
 
Moment Distribution Method For Btech Civil
Moment Distribution Method For Btech CivilMoment Distribution Method For Btech Civil
Moment Distribution Method For Btech Civil
 
Online electricity billing project report..pdf
Online electricity billing project report..pdfOnline electricity billing project report..pdf
Online electricity billing project report..pdf
 
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
 
PE 459 LECTURE 2- natural gas basic concepts and properties
PE 459 LECTURE 2- natural gas basic concepts and propertiesPE 459 LECTURE 2- natural gas basic concepts and properties
PE 459 LECTURE 2- natural gas basic concepts and properties
 

Ijetcas14 536

  • 1. International Association of Scientific Innovation and Research (IASIR) (An Association Unifying the Sciences, Engineering, and Applied Research) International Journal of Emerging Technologies in Computational and Applied Sciences (IJETCAS) www.iasir.net IJETCAS 14-536; © 2014, IJETCAS All Rights Reserved Page 99 ISSN (Print): 2279-0047 ISSN (Online): 2279-0055 Estimation of Inputs for a Desired Output of a Cooperative and Supportive Neural Network 1P. Raja Sekhara Rao, 2K. Venkata Ratnam, 3P.Lalitha 1Department of Mathematics, Government Polytechnic, Addanki - 523 201, Prakasam, A.P., INDIA. 2Department of Mathematics, Birla Institute of Technology and Science-Pilani, Hyderabad campus, Jawahar Nagar, Hyderabad-500078, INDIA. 3Department of Mathematics, St. Francis College for Women, Begumpet, Hyderabad, INDIA. __________________________________________________________________________________________ Abstract: In this paper a cooperative and supportive neural network involving time delays is considered. External inputs to the network are allowed to vary with respect to time. Asymptotic behavior of solutions of network system with variable inputs is studied with respect to its counterpart of constant inputs. With suitable restrictions on the inputs, it is noticed that solution of the network may be made to approach a pre-specified output. Keywords: Co-operative and Supportive Neural Network, Variable Inputs, Desired Output, Convergence. __________________________________________________________________________________________ I. Introduction This paper deals with the study of influence of time varying exogenous inputs on a cooperative and supportive neural network. A model of a cooperative and supportive network (CSNN, for short) is introduced by Sree Hari Rao and Raja Sekhara Rao [9]. It takes into account the collective capabilities of neurons involved with tasks divided and distributed to sub networks of neurons. Applications to such networks are many, for example, in industrial information management (hierarchical systems) which involve distribution and monitoring of various tasks. They are also useful in classification and clustering problems, data mining and financial engineering [6,7,8]. They are also utilized for parameter estimation of auto regressive signals and to decompose complex classification tasks into simpler subtasks and solve then. In a recent paper [11], the authors considered time delays in transmission of information from sub-networks to main one as well as in processing of information in sub-network itself (before transmission of information to main network). Qualitative properties of solutions of the system are studied. Sufficient conditions for global asymptotic stability of equilibrium pattern of the system are established even in the presence of time delays. In the present paper, we wish to consider the CSNN model of [11] with time delays to study the influence of time varying inputs on the system. The motivation for this study stems from the observations of [10] that the applicability a neural network may be increased by the choice of inputs and inputs play a key role in attaining desired outputs. Proper choice of could be an alternative for modifying the neural network for each application and existing neural network may be utilized for different tasks, thus. Besides this, the presence of time varying inputs make the system non autonomous and the study enriches the literature. Mathematical studies of neural networks have been concentrated on stability of equilibrium patterns. Equilibria are stationary solutions of the system and correspond to memory states of the network. Stability of an equilibrium implies a recall of memory state. Thus, such stability analysis of neural networks is confined to recall of memories only and we may not reach the desired output for which the network is intended. In the present study, we deviate from this recall of memories but look for ways of reaching a desired solution. An attempt is made in [9] to explain briefly the influence of variable inputs on asymptotic nature of solutions of CSNN model. The present study extends this work. We concentrate on the interplay between the inputs and outputs of the network. For this, several results are established for estimation or restriction of inputs for getting a desired or pre-specified output and understand the behavior of solutions in the presence of variable inputs. The work also extends the study of [10] carried out for BAM networks. As remarked in [10], convergence to a desired output for a given output explained here should not be confused with convergence of output function of the network. Results are available in literature which consider time varying inputs in various directions [1- 3,5,12] but our emphasis here is on utilization of these inputs to make solutions of system approach an a priori value of output. We reiterate that this is not yet another usual study on qualitative behavior of solutions of the system under the influence of variable inputs. The paper is organized as follows. In Section 2, the model under consideration is explained. Asymptotic behavior of solutions and their relation with the solutions of corresponding system with constant inputs are discussed. Section 3 deals with the input-output trade-off. Estimates on inputs are provided for approaching a desired, preset output for the network. A discussion follows in Section 4.
  • 2. P. Raja Sekhara Rao et al., International Journal of Emerging Technologies in Computational and Applied Sciences, 9(1), June-August, 2014,pp. 99-105 IJETCAS 14-536; © 2014, IJETCAS All Rights Reserved Page 100 II. The Model and Asymptotic Behavior The following model is considered in [11], , In (2.1), , i=1,2,...,n denote a typical neuron in neural field X and denote a subgroup of neurons in another neuronal field Y and are attached to . may be considered to form the main group of neurons which are required to perform the task. constitute a subgroup of neurons attached to each to which assigns some of its task. support, coordinate and cooperate with in completing the task. and are positive constants known as passive decay rates of neurons and respectively. and are the synaptic connection weights and all these are assumed to be real or complex constants. denotes the rate of distribution of information between and . The weight connections connect the i th neuron in one neuronal field to the j th neuron in another neuronal field. The functions and are the neuronal output response functions and are more commonly known as the signal functions. The parameter signifies the time delay in transmission of information from sub network neuron to main network neuron . The delay in second equation represents the processing delays in the subsystems. and are exogenous inputs which are assumed to be constants in [11]. For more details of the terms and design of the CSNN, readers are referred to [9]. Introducing the time variables and (t), , in place of constant inputs and into the system (2.1), we get The following initial functions are assumed for the system (2.2). for , (2.3) where are continuous, bounded functions on and We assume that the response functions and satisfy conditions (2.4) (2.5) (2.6) where , , and are positive constants. Under the conditions (2.4)-(2.6) on the response functions and with and bounded, continuous functions on ) it is not difficult to see that the system (2.2) possesses unique solutions that are continuable in their maximal intervals of existence ([11]). Since (2.2) is non-autonomous it may not possess equilibrium patterns(constant solutions). A solution of (2.1) or (2.2) is denoted by where , throughout. Therefore, we study the asymptotic behavior of its solutions. We recall from [10] that two solutions and of the system (2.2) are asymptotically near if . In the following, we present results on asymptotic nearness of solutions of (2.2). Our first result is Theorem 2.1: For any pair of solutions and of (2.2), we have provided holds, where (2.7)
  • 3. P. Raja Sekhara Rao et al., International Journal of Emerging Technologies in Computational and Applied Sciences, 9(1), June-August, 2014,pp. 99-105 IJETCAS 14-536; © 2014, IJETCAS All Rights Reserved Page 101 Proof: Consider the functional, Along the solutions of (2.2), the upper right derivative of V is given by using (2.4)-(2.6). . Integrating both sides with respect to t, Thus, V (t) is bounded on and for . But and are also bounded on . Hence, it follows that their derivatives are also bounded on
  • 4. P. Raja Sekhara Rao et al., International Journal of Emerging Technologies in Computational and Applied Sciences, 9(1), June-August, 2014,pp. 99-105 IJETCAS 14-536; © 2014, IJETCAS All Rights Reserved Page 102 Therefore, and , are uniformly continuous on . Thus, we may conclude that and (e.g., [4]). This concludes the proof. The following result provides conditions under which all solutions of (2.2) are asymptotic to the solutions of (2.1), which shows that for a proper choice of input functions the stability of system (2.1) is not altered by the presence of time dependent inputs. Theorem 2.2: Assume that the parametric conditions (2.7) hold. Further, let inputs satisfy where . Then for any solutions (x, y) of (2.2) and Proof: To establish this, we employ the same functional as in Theorem 2.1, that is, Proceeding as in Theorem 2.1 we get after a simplification and rearrangement + Rest of the argument is same as that of Theorems 2.1, and hence, omitted. Thus, the conclusion follows. We now recall from [11] that the system (2.1) has a unique equilibrium pattern any set of input vectors provided the parameters satisfy, Then we have, Corollary 2.3: Assume that all the hypotheses of Theorem 2.2 are satisfied. Further if (2.1) possesses equilibrium pattern then all solutions (x, y) of (2.2) approach Proof: The result obviously follows form the observation that the equilibrium pattern is also a solution of (2.1) and the choice in Theorem 2.2. The following example illustrates the above results. Example 2.4: Consider the following system having two neurons in X each supported by two neurons in Y involving time delays as given by . Choose and . Then . Clearly conditions for both the existence of unique equilibrium and its stability are satisfied for any pair of constant inputs .
  • 5. P. Raja Sekhara Rao et al., International Journal of Emerging Technologies in Computational and Applied Sciences, 9(1), June-August, 2014,pp. 99-105 IJETCAS 14-536; © 2014, IJETCAS All Rights Reserved Page 103 Now choose It is easy to see that since the condition holds, and we have (i). Conditions of Theorem 2.1 are satisfied and all solutions of the system are asymptotic to each other. (ii). Conditions of Corollary 2.3 are satisfied and all solutions of the system approach the equilibrium pattern of corresponding system with constant inputs. III. Estimations on Inputs for a Pre-specified Output In this section, we try to estimate our inputs, depending on the output given, that help the solutions to approach the given output. For an easy understanding of the concept, we avoid the complicated notation. We, therefore, rearrange our system (2.2) suitably. We use the following notation For , (2.2) may be represented as (3.1) in which We assume that is the desired output of the network. Further that both and are fixed with respect to t and are arbitrarily chosen. We now arrange (3.1) as (3.2) Conditions on response functions (2.2) to (2.6) may be modified as and for some || denoting appropriate norm. We have Theorem 3.1. Assume that the parameters of the system and the response functions satisfy the condition For an arbitrarily chosen output , the solutions of system (3.1) converge to provided the inputs satisfy either of the conditions or Proof . Then the upper right derivative of V along the solutions of (3.1), using (3.2), we have – – – This gives rise to Rest of the argument is similar to that of Theorem 2.1 and invoking the condition (i) on inputs. Again, it is easy to see from the last inequality above that for large t using conditions (ii) on and . Hence, in either of the cases, follows. The proof is complete. The following example illustrates the effectiveness of this result.
  • 6. P. Raja Sekhara Rao et al., International Journal of Emerging Technologies in Computational and Applied Sciences, 9(1), June-August, 2014,pp. 99-105 IJETCAS 14-536; © 2014, IJETCAS All Rights Reserved Page 104 Example 3.2: Consider Now choose and Clearly, for given we have all the conditions of Theorem 3.1 are satisfied, and hence, for sufficiently large t. We shall now consider the time delay system corresponding to (2.2), (3.3) As has been done earlier for a given , we write (3.3) as (3.4) We have Theorem 3.3: Assume that the parameters of the system and the response functions satisfy the condition For an arbitrarily chosen output , solutions of system (3.3) converge to provided the inputs satisfy the condition Proof: Employing the functional using (3.4) and proceeding as in Theorem 2.1 and Theorem 3.1, the conclusion follows. Since the conditions on parameters and input functions in Theorem 3.1 and 3.3 are the same, it may imply that delays have no effect on convergence here Example 3.4: Consider Now choose and Clearly, for given we have all the conditions of Theorem 3.1 are satisfied, and hence, for sufficiently large t. Remark 3.5: Now consider the system (3.5) are constant inputs, given . It is easy to observe that is an equilibrium pattern of (3.5). Then from Corollary 2.4, we have, solutions of (3.3) approach whenever the variable inputs of (3.3) are well near those of (3.5) as specified in Corollary 2.4. Thus, by varying the external inputs of the system in the parameter space defined by as specified by Theorems 3.1 and 3.3, the solutions of the network approaches pre-specified output IV. Discussion In this article, we have extended the concept of approaching a desired output of a given network by suitable selection of inputs based on the given output for a cooperative and supportive neural network that was studied earlier for a BAM network ([10]). With the help of suitable Lyapunov functionals, results are established for asymptotic nearness and boundedness of solutions of the system also. It is noticed that inputs define a new space of equilibria for the network while they run through a space defined by output parameters. This way memory states of brain that are usually ignored by constant inputs may be recalled by varying the inputs to brain appropriately. Since the input-output relation is not direct but includes system parameters and functional responses, dynamics of entire system are involved in this process. It is hoped that this concept helps in utilizing the same network for different applications without altering its architecture. This shows how designed structures may be made emergent structures which are adaptive and flexible. Since the results hold good for all time delays (delay independent criteria) the results are applicable to delay-free case as well, i.e., models of [9].
  • 7. P. Raja Sekhara Rao et al., International Journal of Emerging Technologies in Computational and Applied Sciences, 9(1), June-August, 2014,pp. 99-105 IJETCAS 14-536; © 2014, IJETCAS All Rights Reserved Page 105 V. References [1] H. Bereketoglu and I. Gyori, Global asymptotic stability in a nonautonomous Lotka-Volterra type system with infinite delay, Journal of Mathematical Analysis and Applications, 210(1997), 279-291. [2] Q.X. Dong, K. Matsui and X.K. Huang, Existence and stability of periodic solutions for Hopfield neural network equations with periodic input, Nonlinear Analysis, 49(2002), 471-479. [3] M. Forti, P. Nistri and D. Papini, Global exponential stability and global convergence in finite time of delayed neural networks with infinite gain, IEEE TNN 16(6): 1449-1463, 2005. [4] K. Gopalsamy and Xue-Zhong He, Delay-independent stability in bidirectional associative memory networks, IEEE TNN, 5(1994), 998-1002. [5] S. Hu and D. Liu, On global output convergence of a class of recurrent neural networks with time varying inputs, 18(2005), 171- 178. [6] B. Kosko, “Neural Networks and Fuzzy Systems - A Dynamical Systems Approach to Machine Intelligence", Prentice-Hall of India, New Delhi, 1994. [7] F.-L. Luo, R. Unbehauen, Applied Neural Networks for Signal Processing, Cambridge Univ. Press, Cambridge, UK, 1997. [8] B.B. Misra and S. Dehuri, Functional link artificial neural network for classification task in data mining, J. Computer Science, 3(12), 2007, 948-955. [9] V. Sree Hari Rao and P. Raja Sekhara Rao, Cooperative and Supportive Neural Networks, Physics Letters A 371 (2007) 101–110. [10] V. Sree Hari Rao and P. Raja Sekhara Rao, Time Varying Stimulations to Simple Neural Networks and Convergence to Desired Outputs, Communicated. [11] P. Raja Sekhara Rao, K.Venkata Ratnam and P. Lalitha, Delay Independent Stability of Co-operative and Supportive Neural Networks, Communicated. [12] Zhang Yi, J.C. Lv and L. Zhang, Output convergence analysis for a class of delayed recurrent neural networks with time varying inputs, IEEE TSMC, 36(1), 87-95, 2006.